Although much is known about client prognostic indicators of substance abuse treatment outcome, less is known about how treatment impacts outcome. In particular, there is a need for better information on how to allocate inpatient and outpatient care in relation to the severity of substance abuse, psychiatric disorders and life context factors. Such research holds promise for improving the effectiveness and reducing the overall costs of substance abuse treatment. The primary aims of this project are to: (1) assess the cost-effectiveness of residential and outpatient treatment for substance abuse; (2) document treatment and posttreatment service utilization; (3) evaluate three theory-based patient-treatment matching hypotheses; and (4) assess the patient- treatment matching utility of an established patient assignment system. A total of 600 participants with drug abuse problems will be randomly assigned to residential treatment (N = 200), outpatient treatment (N = 200), or to a usual care group in which participants receive a program placement based on the American Society of Addiction Medicine's patient assignment criteria, followed by treatment in a flexible continuum of care (N = 200). These clients will be assessed on a wide variety of health and psychosocial measures prior to treatment and at six- and twelve-months after discharge from treatment. Service utilization and cost data, as well as data on treatment participation, intermediate outcomes, and life context factors, will also be collected and analyzed. The proposed study will overcome the weaknesses of previous studies in several ways: (1) study clients will be randomly assigned to alternative treatment settings; (2) multiple outcomes will be measured in a longitudinal design and substance use will be verified using urine screens and collateral reports; (3) services utilization will be assessed using a comprehensive standard service taxonomy and individual client costs will be determined using valid and reliable methods; and (4) multivariate analyses will be undertaken to examine the processes of treatment and to test a priori hypotheses concerning differential effects of treatment setting for different types of clients. The cost and outcome data will be analyzed using a number of multivariate statistical analyses, including the Theta methodology, which is designed for analyzing outcomes and costs together. Structural equation modeling, regression, and analysis of covariance will also be used in cost- effectiveness analyses. Comparisons of treatment outcomes and tests of matching hypotheses will be conducted using selection modeling, analysis of covariance, and classification and regression tree analyses. Mediational models of treatment processes that link treatment characteristics to intermediate outcomes and final outcomes will be assessed using a path analytic strategy.